Results from analysis of cities' climate plans

Look in notebooks for the analysis

Contains code for running OCR libraries on PDFS to get text outputs.

text-preprocessing.ipynb Run this first to get serialized features used for the other notebooks
Keyterm-based topic analysis.ipynb uses word2vec and tfidf to construct emissions source lexicons, and uses factor analysis to show how cities are taking different climate strategies
Net Zero Ambition Regression.ipynb uses the text features and labels from human review of which plans have a concrete goal of reaching net zero, and performs logistic regression to find significant terms that distinguish ambitious plans from non-ambitious plans
similarity_valuation.ipynb applies pairwise similarity of text features to identify how regions and coordinator networks influence climate plans' language